QGIS enacts moratorium on AI contributions
- ID
- 34533
- Status
- summarized
- Published
- 12 Oct 2026, 7:22 PM
- Fetched
- 13 Oct 2026, 5:58 AM
- Provider
- Hacker News
- Category
- dev-community
- Original URL
- https://github.com/qgis/QGIS-Enhancement-Proposals/pull/387
- Source URL
- https://hnrss.org/best
Summary
- Score
- 7.0
- Created
- 13 Oct 2026, 5:59 AM
- Tags
- Audience
- developersvibe_coderssaas_founders
What happened
QGIS merged QEP428 into its QGIS-Enhancement-Proposals repo on 7 Oct 2026, formally replacing its previous 'human in the loop' AI policy with a strict 'no AI contributions permitted' rule. The pull request drew 68 thumbs-up, 19 thumbs-down, 9 heart and 5 eyes reactions, and the Hacker News thread hit 352 points and 386 comments. Maintainer haubourg justified the move by saying AI-generated contributions put 'massive and unbearable pressure' on maintainers, that the AI submission stream exceeds what human review can absorb, and that QGIS's organisational trust rests on a human contributor network that can be hired for support, training and features.
Why it matters
If you use Copilot, Cursor or Claude Code to draft patches for QGIS, those contributions are now rejected outright rather than merely needing human review — the practical path for a fix you need is to reimplement it by hand or contract a contributor, since the text says organisations value QGIS partly because they can hire contributors for support and feature work. It is also a concrete precedent for anyone shipping or maintaining open source: the stated cost driver is maintainer review bandwidth, not code quality in the abstract, so projects with thinner maintainer pools are the ones most likely to copy this. No Malaysia-specific element appears in the source text.
Discussion angle
The reaction split (68 up vs 19 down) shows the community is not unanimous — is an outright ban the right call, or would a disclosure-plus-accountability tier be more workable given maintainers say the AI PR stream already exceeds review capacity?